PCA Segmentation of Multivariate Time Series and Its Application in the Condition Identification of Natural Gas Dehydration Unit
Received:July 05, 2019  Revised:September 05, 2019
View Full Text  View/Add Comment  Download reader
DOI:10.7643/issn.1672-9242.2020.04.014
KeyWord:multivariate time series  PCA  sequence segmentation  parameter clustering  separation of working conditions
                       
AuthorInstitution
SONG Wei Southwest Oil and Gasfield Company Chongqing Gas District, Chongqing , China
XIONG Wei Southwest Oil and Gasfield Company Chongqing Gas District, Chongqing , China
DONG Sha-sha Southwest Oil and Gasfield Company Chongqing Gas District, Chongqing , China
TAN Jian Southwest Oil and Gasfield Company Chongqing Gas District, Chongqing , China
PENG Bo Southwest Oil and Gasfield Company Chongqing Gas District, Chongqing , China
WU Jiao Southwest Oil and Gasfield Company Chongqing Gas District, Chongqing , China
LIANG Tian-you a.State Key Laboratory of Mechanical Transmissions, b.School of Mechanical Engineering, Chongqing University, Chongqing , China
YIN Ai-jun a.State Key Laboratory of Mechanical Transmissions, b.School of Mechanical Engineering, Chongqing University, Chongqing , China
Hits:
Download times:
Abstract:
      The work aims to establish the equipment state prediction and evaluation models in the different conditions. The PCA method was used to cut the multivariate time series data, and the segmented data segments were clustered and merged according to the density-based method and the defined distance to obtain the time series under different conditions. According to the operating conditions of the dehydration unit, the multivariate time series corresponding to reboiler characteristics were divided into different data segments. The parameter data segmentation of the reboiler of the dehydration unit is effectively realized, and different working conditions are identified.
Close